The standard treatment regimen for active tuberculosis (TB) is >40 years old and has suboptimal effectiveness, because it requires daily drug adherence for ?6 months, long after symptoms have resolved. Now, for the first time in decades, the TB drug pipeline is full of new, highly potent and possibly treatment- shortening agents nearing readiness for phase 2/3 clinical trials. Unfortunately, evaluating these drugs poses major challenges, because of the large sample sizes (~2500 patients) and long follow-up periods (~30 months) required to show their non-inferiority to the standard regimen in preventing treatment failure and relapse. Although two-month mycobacterial culture conversion is currently used as a surrogate marker, it has numerous disadvantages, such as the need to collect sputum from patients who are no longer coughing, transport it to specialized reference laboratories, and wait weeks for a dichotomous result which has poor statistical power. The objective of this proposal is to develop a quantitative, blood-based, TB biomarker assay using a proteomic approach that targets human exosomes. Exosomes are vesicular nanoparticles released by host cells during active infection. Consistent with their postulated role in immune signaling, exosomes contain high concentrations of M. tuberculosis (Mtb) proteins, which seem to decrease with clearance of Mtb from infected mice. Thus, we hypothesize that exosome proteins can serve as suitable biomarkers for predicting microbiologic response to TB treatment among groups randomized to different drug regimens in clinical trials, as well as among individuals on standard regimens. A non-sputum-dependent, non-growth-based biomarker assay that could be applied early in treatment in either or both of these scenarios could transform the pace and scope of TB drug development and global TB control. This assay would also be useful for monitoring treatment of paucibacillary disease as is common in childhood TB, extrapulmonary TB, and HIV-TB co-infection. To achieve this goal, we have assembled a multidisciplinary team of TB researchers with expertise in exosome biology, proteomics, diagnostics, novel prediction methods, and clinical trials. We will build on the existing infrastructure of my K23 mentor's NIH-funded cohort study of pneumonia by collecting serial serum specimens from patients followed through the initiation phase of treatment for culture-confirmed pulmonary TB. We will use these data and samples to 1) optimize methods for processing serum exosomes for Mtb- biomarker assays; 2) evaluate the kinetics of a serologic, exosome-based Mtb-biomarker panel in response to treatment, and determine the optimal sampling frame for assessing that response. Upon completing these proof-of-principle studies, we expect to have the preliminary data to justify larger clinical studies to determine the accuracy of the marker in predicting treatment response in test and validation cohorts; to further develop and validate a clinical testing platform for these markers; and ultimately to evaluate the resultin refined assay as a surrogate marker in Phase 3 clinical trial banks currently under development. PUBLIC HEALTH RELEVANCE: Because current tuberculosis (TB) treatments are long and hard to tolerate, many people around the world fail treatment and die. New drugs for TB patients are now in development but will take years to come to market, unless we can create tests to predict early in treatment which drugs are most effective. Such tests would speed new drug development and improve individual care by predict when treatments are likely to fail. We plan to develop such a test using human exosomes, newly identified bloodstream particles which carry TB proteins.